Prediction of Equilibrium Flow and Bed Load Transport in a Curved Bend

A recently developed fully non-hydrostatic model to predict flow, sediment transport and equilibrium bathymetry using RANS models without wall functions and with capabilities to account for bed roughness effects is used to predict the equilibrium flow and sediment transport in a 140° curved channel bend studied experimentally by Struiksma under incoming steady-flow conditions. Additionally, a bed load transport module based on Engelund and Hansen total bed load formula is implemented and tested. In the experiment, the bed load was observed to be dominant, so the suspended sediment component is considered negligible in the simulations. The streamwise variation of the relative water depths at several representative sections are compared with the experimental data. It is found that the best agreement is obtained when the total load (which in the test case considered is equal to the bed load) is estimated using Engelund and Hansen formula with both k-ω and Spalart-Allmaras models. The k-ω predictions, using the originally implemented non-equilibrium bed load transport model based on van Rijn and Wu et al. methodology, are found to overpredict the water depths in the first half of the bend. The main reason for the underperformance of the van Rijn model is its inability to predict the experimentally measured bed load in the initial straight part of the channel.